Pressure-dependent Leakage Detection Research And Application

From the perspective of economy，the leakage results in direct economic lossesto the water utility,it increases cost due to the lost revenue; From the perspective ofresources,the leakage meas the waste of water resources. A total cost of$14.6billion has been estimated for water loss around the world. But when it comes toChina,the problem may be more serious as the research on pipe network started toolate and the construction and maintenance of the pipe network is even morebackward,every year the total water consumption for leakage is already60billion tonnes. In such a severe situation, the research on leakage is necessary.Firstly,this paper gives a brief introduce of the particle swarmoptimization.As an emerging swarm intelligence optimization algorithm, PSO wasfirstly proposed by two American scientists Kennedy and Barnhart in1995. In thispaper, the PSO,GAand B-P are compared through theoretical analysis and testfunction.When the optimal solution for the test function is zero and the iterations is100、200, PSO can find the optimal solution more quickly and the optimal solutionit find can be more closer to zero,which fully verified its strong ability of practicalapplication,what’s more,it provides a powerful tool for the following pumpchecking and leakage model of optimal solution.Secondly,the build process of the hydraulic model in city A was introduced inthe paper. The hydraulic model is the basic to build a leakage model,and the writerintroduces a complete process of building a hydraulic model in details from fouraspects,which are the establishment of the topology of the network,the distributionof the water,pump and model calibration. Model checking is the crux to verifywhether the established models is proper. Only making error in allowed range canput the model into application.Thirdly,in this paper,it adopts the principle of pressure-dependent leakagedetection to locating the leakage of pipe network. Leakage were positivelycorrelated with the pressure, the greater the pressure, the greater the loss,based onthis,the leakage can be seen as a emitter flow, and each node be given a emittercoefficient. If the emitter coefficient is greater than zero, it means the node is theleakage hospots. The greater the coefficient, the greater the loss. By using the particle swarm algorithm for the optimization of simple model, it verifys thefeasibility of pressure-dependent leakage detection, and analyze the influence ofthe amount of pressure monitor for the locating of the leakage.Finally, this paper applys pressure-dependent leakage detection in actual pipenetwork in city A. Based on microscopic hydraulic model of city A, leakage modelis established using one submodel of city A. And by using PSO to solve the leakagemodel,the author locates the leakage out of the pipe network accutately.The furthestdistance between the simulated leakage node and the actual leakage node is467mand the closest distance is125m。...